search/documents/physical_htm.php

   1 <?php
   2 /**
   3 * Global Search Engine for Moodle
   4 *
   5 * @package search
   6 * @category core
   7 * @subpackage document_wrappers
   8 * @author Valery Fremaux [valery.fremaux@club-internet.fr] > 1.8
   9 * @date 2008/03/31
  10 * @license http://www.gnu.org/copyleft/gpl.html GNU Public License
  11 * @version revised for Moodle 2.0
  12 *
  13 * this is a format handler for getting text out of a proprietary binary format
  14 * so it can be indexed by Lucene search engine
  15 */
  16
  17 /**
  18 * @param object $resource
  19 * @param string $directfile if the resource is given as a direct file path, use it as reference to the file
  20 * @uses $CFG
  21 */
  22 function get_text_for_indexing_htm(&$resource, $directfile = ''){
  23     global $CFG;
  24
  25     // SECURITY : do not allow non admin execute anything on system !!
  26     if (!has_capability('moodle/site:config', get_context_instance(CONTEXT_SYSTEM))) return;
  27
  28     // just get text
  29     if ($directfile == ''){
  30         $text = implode('', file("{$CFG->dataroot}/{$resource->course}/{$resource->reference}"));
  31     } else {
  32         $text = implode('', file("{$CFG->dataroot}/{$directfile}"));
  33     }
  34
  35     // extract keywords and other interesting meta information and put it back as real content for indexing
  36     if (preg_match('/(.*)<meta ([^>]*)>(.*)/is', $text, $matches)){
  37         $prefix = $matches[1];
  38         $meta_attributes = $matches[2];
  39         $suffix = $matches[3];
  40         if (preg_match('/name="(keywords|description)"/i', $meta_attributes)){
  41             preg_match('/content="([^"]+)"/i', $meta_attributes, $matches);
  42             $text = $prefix.' '.$matches[1].' '.$suffix;
  43         }
  44     }
  45     // brutally filters all html tags
  46     $text = preg_replace("/<[^>]*>/", '', $text);
  47     $text = preg_replace("/<!--[^>]*-->/", '', $text);
  48     $text = html_entity_decode($text, ENT_COMPAT, 'UTF-8');
  49     $text = mb_convert_encoding($text, 'UTF-8', 'auto');
  50
  51     /*
  52     * debug code for tracing input
  53     echo "<hr/>";
  54     $FILE = fopen("filetrace.log", 'w');
  55     fwrite($FILE, $text);
  56     fclose($FILE);
  57     echo "<hr/>";
  58     */
  59
  60     if (!empty($CFG->block_search_limit_index_body)){
  61         $text = shorten_text($text, $CFG->block_search_limit_index_body);
  62     }
  63     return $text;
  64 }
  65 ?>